Employee Compensation - Theory, Practice & Evidence

8/6/2019 Employee Compensation - Theory, Practice & Evidence

http://slidepdf.com/reader/full/employee-compensation-theory-practice-evidence 1/21

Cornell University ILR School

DigitalCommons@ILRCAHRS Working Paper Series Center for Advanced Human Resource Studies(CAHRS)5-1-1995

Employee Compensation: Theory,Practice, andEvidenceBarry A. GerhartCornell University

Harvey B. Minkoff TRW Corporation

Ray N. OlsenTRW CorporationThis Article is brought to you for free and open access by the Center for Advanced Human Resource Studies (CAHRS) atDigitalCommons@ILR. Ithas been accepted for inclusion in CAHRS Working Paper Series by an authorized administrator of DigitalCommons@ILR. For more information,please contact [email protected], Barry A.; Minkoff, Harvey B. ; and Olsen, Ray N. , "Employee Compensation: Theory, Practice, andEvidence" (1995).CAHRS Working Paper Series. Paper 194.http://digitalcommons.ilr.cornell.edu/cahrswp/194

W O R K I N G P A P E R S E R I

E S

Employee Compensation:Theory, Practice, and EvidenceBarry GerhartHarvey Minkoff Ray OlsenWorking Paper 9 5 ² 0 4CAHRS / Cornell University187 Ives HallIthaca, NY 14853-3901 USATel. 607 255-9358www.ilr.cornell.edu/CAHRS/

Advancing the World of WorkEmployee Compensation WP 95-04Page 1



Employee Compensation: Theory, Practice, and EvidenceBarry GerhartCenter for Advanced Human Resource StudiesCornell UniversityHarvey B. Minkoff and Ray N. OlsenTRW Corporation

May 1994Working Paper #95-04www.ilr.cornell.edu/cahrs

An updated version of this working paper has been published in Handbook of Human ResourceManagement, Ferris, Rosen and Barnum (Ed.), Chapter (27), 1995.This paper has not undergone formal review or approval of the faculty of the ILR School. It isintended to make results of research, conferences, and projects available to others interested inhuman resource management in preliminary form to encourage discussion and suggestions.Employee Compensation WP 95-04Page 2

INTRODUCTION As organizations continue to face mounting competitive pressures, they seek to do more

with less and do it with better quality. As goals for sales volume, profits, innovation, and qualityare raised, employment growth is often tightly controlled and in many cases, substantial cuts inemployment have been made. To accomplish more with fewer employees calls for effectivemanagement of human resources. Typically, the employee compensation system, the focus of this chapter, plays a major role in efforts to manage human resources better.Employee compensation plays such a key role because it is at the heart of theemployment relationship, being of critical importance to both employees and employers.Employees typically depend on wages, salaries, and so forth to provide a large share of their income and on benefits to provide income and health security. For employers, compensationdecisions influence their cost of doing business and thus, their ability to sell at a competitiveprice in the product market. In addition, compensation decisions influence the employer's abilityto compete for employees in the labor market (attract and retain), as well as their attitudes and

behaviors while with the employer.Employee compensation practices differ across employment units (e.g., organizations,business units, and facilities) on several dimensions (Gerhart & Milkovich, 1990, 1992; Gerhart,Milkovich, & Murray, 1992). The focus of the employee compensation literature has been ondefining these dimensions, understanding why organizations differ on them (determinants), andassessing whether such differences have consequences for employee attitudes and behaviors,and for organizational effectiveness. In the following discussion, we briefly describe the basicdimensions of compensation and summarize some of the key theories used to explain theconsequences of different compensation decisions. A discussion of pay determinants can befound in Gerhart and Milkovich (1990, 1992).STRATEGIC PAY DIMENSIONSPay practices vary significantly across employing units and to some degree, across jobs.

We discuss the form, level, structure, mix, and administration of payment systems (Gerhart &Milkovich, 1992; Heneman & Schwab, 1979; Milkovich & Newman, 1993).First, pay can be in the form of cash or benefits (e.g., health care, retirement, paidvacation). On average, about 70 percent of payments to U.S. employees are in the form of cash, leaving 30 percent in the form of noncash and deferred cash benefits (Noe, Hollenbeck,Gerhart, & Wright, 1994). Health care has been the fastest growing benefit, and most employersdescribe the challenge of controlling this cost while providing quality coverage as one of their top human resource management challenges.Employee Compensation WP 95-04



Page 3

Second, both benefits and cash compensation can be described in terms of their level(how much). Most organizations use one or more market pay surveys to help determine whatother organizations pay specific jobs in making their own pay level decisions. More broadly, totallabor costs are a function of both compensation cost per employee and total employeeheadcount. Therefore, to assess competitiveness in the product market, organizations should

not focus only on pay levels. They should compare total labor costs, and better yet, they shouldcompare with other organizations the sort of return (or productivity) they receive in terms of profits, sales, and so forth for each dollar spent on labor costs. The now commonannouncements of major reductions in force attest to the importance of controlling labor costs.Such decisions are also sometimes driven by comparisons of revenue or profits per employee,or the ratio of sales or profits to labor costs.Labor costs and productivity are also key factors in decisions about where to locateproduction. Germany's high labor costs have led to what Business Week described as the"Exodus of German Industry." German companies are moving production to lower labor costcountries, such as those in Eastern Europe and the United States.1 BMW recently announced itwould be building vehicles in South Carolina and Mercedes-Benz will produce vehicles in

Alabama. Agreements such as the North American Free Trade Agreement.(NAFTA) and the

General Agreement on Trade and Tariffs will only reinforce the globalization of production.However, contrary to what was heard in the debate over NAFTA, labor costs will not be thedetermining factor in most cases, except perhaps for labor-intensive production. Labor costs asa percentage of total costs is shrinking in many cases, and other factors such as access tomarkets and labor force quality will often be more important. The decision by BMW andMercedes-Benz to build in the United States, not Mexico, is evidence of this.Third, the structure refers to the nature of pay differentials within an employing unit. Howmany steps or grades are in the structure? How big are the pay differentials between differentlevels in the structure? Large organizations often have over 20 such levels, although manyorganizations have recently reduced the number of steps ("delayered"). Are employees at thesame hierarchical level in different parts of the organization (e.g., different product sectors or different occupational groups) paid the same? Yet another aspect of structure is the timing of

payment over employees' careers. Some organizations may bring entry level people in at arelatively high rate of pay, but then provide relatively slow pay growth, while another 1 Labor costs in the United States have been lower- than those of Germany in recent years. However, this differencechanges as currency exchange rates differ. The United States still has the highest purchasing power per capita of anycountry.Employee Compensation WP 95-04Page 4

organization may bring employees in relatively low but offer greater opportunities for promotionand pay growth over time.Fourth, payment systems differ in their mix (how and when cash compensation isdisbursed). Some organizations pay virtually all employees a base salary that is adjustedapproximately once per year through a traditional merit increase program. Merit increasesbecome part of base salary and are supposed to depend on merit (performance), although thereis a widespread belief that most employees get about the same percentage increase, regardlessof their performance. As described below, an increasing number of organizations are usingso-called variable pay or pay at risk, which means that some portion of employees' pay isuncertain and depends on some combination of future business unit or organizationperformance (e.g., profits, stock performance, productivity), group performance, and individualperformance. Specific pay programs that influence pay mix are merit pay, incentive pay,gainsharing, profit sharing, and stock plans (e.g., stock options).Fifth, pay is administered differently in different organizations. The design of pay policies



differs, for example, in terms of who is involved in the process. The roles of human resourcedepartments, line managers, and rank and file employees differ across situations. In someorganizations, line managers may design plans, often with assistance from the humanresources department. Alternatively, human resources takes the lead in other cases.Employees to be covered by a payment system are sometimes involved, and in some cases,may actually design plans for themselves.

Communication is another aspect of administration. The most technically sophisticatedpayment plan can generate desired employee reactions or exactly the opposite. The actualeffect depends on whether the rationale for the payment plan is understood and accepted andwhether employees' perceptions of the facts upon which the rationale is built (e.g., thecompany's financial health, the pay of employees in other jobs or organizations) are the sameas the perceptions of those charged with seeing that the payment plan has the intended effects.We focus in this chapter on cash compensation issues. Benefits warrants a chapter of itsown and discussions are available elsewhere (Beam & McFadden, 1992; Gerhart & Milkovich,1992; Noe, Hollenbeck, Gerhart, & Wright, 1994). Further, our discussion of cash compensationis mostly limited to pay mix issues, an area that has been of great interest to organizations asthey move (or consider a move) to "new" programs such as stock plans for non-executives,gainsharing, and profit sharing. In the remainder of this chapter, we provide a survey of theories

that have been used to study the effects of pay decisions, describe specific pay programs andtheir expected consequences, and review recent empirical evidence on that question.Employee Compensation WP 95-04Page 5

CONSEQUENCES OF PAY DECISIONS: THEORIESTo understand what types of pay systems are most likely to be effective and how their effectiveness differs according to contingency factors such as business strategy, nationalculture, competitive environment, and employee characteristics, we need to have a goodconceptual framework, or theory. In truth, there is as of yet no grand theory of compensationthat takes these contingency factors into account, although recent work by Gomez-Mejia andBalkin (1992) is promising.In examining consequences, we need to recognize that effectiveness is a multi-faceted

concept that could include at a minimum, cost, productivity, innovation, quality, financial, andattitudinal dimensions. Further, the relative importance of these dimensions will vary acrossorganizations and business units.

At the individual level of analysis, theories have been used to show how pay plans canbe used to energize, direct, and control employee behavior. We briefly describe three suchtheories used in research on pay.Reinforcement and Expectancy TheoriesReinforcement theory states that a response followed by a reward is more likely to recur in the future (Thorndike's Law of Effect). The implication for compensation management is thathigh employee performance followed by a monetary reward will make future high performancemore likely. By the same token, high performance not followed by a reward will make it lesslikely in the future. The theory emphasizes the importance of a person actually experiencing the

reward.Like reinforcement theory, expectancy theory (Vroom, 1964) focuses on the link betweenrewards and behaviors (instrumentality perceptions), although it emphasizes expected (rather than experienced) rewards (i.e., incentives). Motivation is also a function of two other factors:expectancy, the perceived link between effort and performance, and valence, the expectedvalue of outcomes (e.g., rewards). Compensation systems differ according to their impact onthese motivational components. Generally speaking, pay systems differ most in their impact oninstrumentality: the perceived link between behaviors and pay, also referred to in the payliterature as "line of sight." Valence of pay outcomes should remain the same under different



pay systems. Expectancy perceptions often have more to do with job design and training thanpay systems.Employee Compensation WP 95-04Page 6

Equity TheoryEquity theory suggests that employee perceptions of what they contribute to the

organization, what they get in return, and how their return-contribution ratio compares to othersinside and outside the organization,' determine how fair they perceive their employmentrelationship to be (Adams, 1963). Perceptions of inequity are expected to cause employees totake actions to restore equity. Unfortunately, some such actions (e.g., quitting or lack of cooperation) may not be helpful to the organization.Two recent empirical studies provide good examples of the types of counterproductivebehaviors that can occur as a result of perceived inequity. In the first study, Greenberg (1990)examined how an organization2 communicated pay cuts to its employees and the effects ontheft rates and perceived equity. Two organization units received 15% across-the-board paycuts. A third unit received no pay cut and served as a control group. The reasons for the paycuts were communicated in different ways to the two pay-cut groups. In the "adequateexplanation" pay-cut group, management provided a significant degree of information to explain

its reasons for the pay cut, and also expressed significant remorse. In contrast, the "inadequateexplanation" group received much less information and no indication of remorse. The controlgroup received no pay cut (and thus no explanation).The control group and the two pay-cut groups began with the same theft rates andequity perceptions. After the pay cut, the theft rate was 54% higher in the adequate explanationgroup than in the control group. However, in the "inadequate explanation" condition, the theftrate was 141% than in the control group. In this case, communication had a large, independenteffect on employees' attitudes and behaviors.Cowherd and Levine (1992) used a sample 102 business units in 41 corporations toexamine whether the size of the pay differential between lower-level employees and topmanagement had any impact on product quality. Cowherd and Levine suggest that individualsoften compare their pay to that of people higher in the organization structure. If lower-level

employees feel inequitably treated, they may seek to reduce their effort to achieve equity.Quality, in their study, was defined as customer perceptions of the quality of goods andservices. They hypothesized that extrarole, or citizenship behaviors, such as freely offering tohelp others, following the spirit rather than letter of rules, and correcting errors that wouldordinarily escape notice, would be less likely when pay differentials between hourly and topmanagerial employees were large. Their results supported this hypothesis, suggesting that2 Employees may use other comparisons standards also, such as their previous or expected future jobs or cost of living.Employee Compensation WP 95-04Page 7

organizations need to take care that they not forget the potential adverse motivationalconsequences of executive pay for the motivation of other employees.Agency Theory

Agency theory, until recently best known in the economics, finance, and law literatures,focuses on the divergent interests and goals of the organization's stakeholders, and the waysthat employee compensation can be used to align these interests and goals (Eisenhardt, 1989;Fama & Jensen, 1983). ownership and management (or control) are typically separate in themodern corporation, unlike the days when the owner and manager were often the same person.With most stockholders far removed from day-to-day operations, so-called agency costs (i.e.,costs that arise from the interests of the principals/owners and their agents/managers notconverging are created. What is best for the agent/manager, may not be best for the owner.



Examples of agency costs include management spending money on perquisites (e.g.,"superfluous" corporate jets) or "empire building" (acquisitions that do not add value to thecompany but may enhance the manager's prestige or pay) rather than seeking to maximizeshareholder wealth (Lambert & Larcker, 1989). In addition, the fact that managers andshareholders may differ in their attitudes toward risk gives rise to agency costs. Shareholderscan diversify their investments (and thus their risks) more easily than managers can diversify

risk in their pay. As a consequence, managers may prefer relatively little risk in their pay (e.g.,high emphasis on base salary,-low emphasis on uncertain bonuses or incentives). Indeed,research shows that managerial compensation in manager-controlled firths is more oftendesigned in this manner (Tosi & Gomez-Mejia, 1989). Agency costs also stem from differencesin decision-making horizons. Especially where managers expect to spend little time in the job or with the organization, they may be more inclined to maximize short-run performance (and pay),perhaps at the expense of long-term success.

Agency theory is also of value in the analysis and design of non-managers'compensation. In this case, the divergence of interests may exist between managers (now inthe role of principals) and their employees (who take on the role of agents). In designing either managerial or non-managerial compensation, the key question is, "How can such agency costsbe minimized?" Agency theory says that the principal must choose a contracting scheme that

helps align the interests of the agent with the principal's own interests (i.e., reduces agencycosts). These contracts can be classified as either behavior oriented (e.g., merit pay) or outcome oriented (e.g., stock options, profit sharing, commissions).

At first blush, outcome-oriented contracts seem to be the obvious solution. If profits arehigh, compensation goes up. If profits go down, compensation goes down. The interests of "theEmployee Compensation WP 95-04Page 8

firm" and employees are aligned. An important drawback, however, is that such contractsincrease the amount of risk borne by the agent. Furthermore, because agents are averse to risk,they may require higher pay (a compensating wage differential) to make up for it.Behavior-based contracts, on the other hand, do not transfer risk to the agent, and thusdo not require a compensating wage differential. However, the principal must be able to monitor

with little cost what the agent has done. Otherwise, the principal must either invest inmonitoring/information or structure the contract so that pay is linked at least partly to outcomes.Which type of contract should an organization use? It depends partly on the followingfactors (Eisenhardt, 1989):jRisk aversion. Risk aversion among agents makes outcome-oriented contracts morecostly.jOutcome uncertainty. Profit is an example of an outcome. Linking pay to profits(outcome-based contract) is more costly to the extent that profits vary and so thereis a risk of low profits.jJob programmability. As jobs become less programmable (i.e., less routine and lessstructured), and more difficult to monitor, outcome-oriented contracts become morelikely. The increasing complexity of organizations and technology makes monitoring

more difficult, and may help explain the growing use of variable pay programs(discussed below), which are examples of outcome-based contracts. Consistent withthis idea, outcome-oriented contracts (e.g., profit sharing and stock plans) are moreprevalent in research and development organizations, where monitoring is especiallydifficult (Milkovich, Gerhart, & Hannon, 1991). Pay levels are also higher, consistentwith the idea that employees must be compensated for sharing more risk.jMeasurable lob outcomes. When outcomes are more measurable, outcome-orientedcontracts are more likely.j Ability to pay. outcome-oriented contracts contribute to higher compensation costs



because of the risk premium.jTradition. A tradition or custom of using (or not using) outcome-oriented contracts willmake such contracts more (or less) likely.Influences on Labor Force CompositionTraditionally, the theories described above have been used to understand how using payto recognize individual contributions can influence the behaviors and attitudes of current

employees, whereas pay level and benefits have been seen as a way to influence so-calledmembership behaviors: decisions about whether to join or remain with the organization.Employee Compensation WP 95-04Page 9

However, there is increasing recognition that individual pay programs may also have an effecton the nature and composition of an organization's work force (Milkovich & Wigdor, 1991;Gerhart & Milkovich, 1992). For example, it is possible that an organization that links pay toperformance may attract more high performers than an organization that does not link the two.There may be a similar effect with respect to job retention.Breaking things down further, perhaps organizations that link pay to individualperformance are more likely to attract individualistic types of employees, while organizationsrelying more heavily on team rewards are more likely to attract more team-oriented employees.

Although there is no concrete evidence of this yet, it has been found that different pay systemsattract different people depending on their personality traits and values (Bretz, Ash, & Dreher,1989; Judge & Bretz, 1992). The implication is that the design of compensation programs needsto be carefully coordinated with the business and human resource strategy.StrategyMoving from the individual level of analysis to the business unit and corporate level,there are theories of what corporate and pay strategies fit best together. Stage in the product lifecycle (Ellig, 1981) and the degree and process of diversification (Kerr, 1985) have been raisedas contingency factors in the design of pay strategies (Milkovich, 1988). Briefly, organizations(or probably more precisely, business units) may go through growth, maintenance, and declinestages, each of which calls for a different compensation strategy. For example, in the growthstage, it was recommended that there be substantial pay at risk to provide high upside earnings

potential (e.g., using stock plans) to spur innovation, growth, and risk-taking, combined with lowfixed costs (base salary and benefits) to preserve scarce capital for investment. In themaintenance and decline stages, there would be less emphasis on pay at risk (except perhapsfor more short term focused plans), and more dollars allocated to base salary and benefits.The literature on diversification and pay strategy suggests that single product firths andunrelated product firms (e.g., conglomerates) have more pay at risk than related product firms,and pay is more decentralized and tied to business unit rather than corporate performance inthe unrelated product firms. This flexibility makes sense where each business unit hasindependent goals, and there is little need for coordination and thus for consistency in paypractices. From an agency theory point of view, it may be more necessary to rely onoutcome-oriented contracts in unrelated products firms because the market-specific expertise isconcentrated in the business units, making it difficult for corporate headquarters to make

evaluations using behavior-oriented contracts. Finally, an unrelated firm that is a result of mergers and acquisitions is more likely to have the flexibility and pay linked to unit performanceEmployee Compensation WP 95-04Page 10

than an unrelated firm that is the result of internal growth because there is often moreinterdependence and interaction in the latter case (Kerr, 1985).Gomez-Mejia and Balkin (Gomez-Mejia & Balkin, 1992; Gomez-Mejia, 1992) havesummarized much of the research on these questions, and provided some of the first tests of whether firms that choose pay strategies consistent with the above frameworks actually perform



better. The answer seems to be "yes." They have provided propositions about which types of pay practices are likely to be most effective based on various strategy frameworks. For example, the Miles and Snow (1978) model classifies business units as defenders (stablemarkets, focus on efficiency), prospectors (focus on new markets and technologies), andanalyzers, which have elements of both defenders and prospectors. According to Gomez-Mejiaand Balkin, variable pay, for example, should be higher in the prospector business units than in

the defender business units. Table 1 shows other proposed differences.Table 1. Matching Organization Strategy and Pay StrategyBusiness Unit StrategyDefenders ProspectorsPay Strategy DimensionsRisk Sharing (Variable Pay) Low HighTime Orientation Short-term Long-termPay Level (short run) Above market Below marketPay Level (long run potential) Below market Above MarketBenefits Level Above market Below marketCentralization of Pay Decisions Centralized DecentralizedPay Unit of Analysis Job Skills

Source: Adapted from Gomez-Mejia, L.R. & Balkin, D.B. (1992). Compensation, organizational strategy, andfirm performance, Appendix 4b.Employee Compensation WP 95-04Page 11

SummaryReinforcement, expectancy, and agency theories all focus on the fact thatbehavior-reward contingencies can shape behaviors. However, agency theory is of particular value in studying variable pay because of its emphasis on the risk-rewardtrade-off, an issue that needs close attention when considering variable pay plans, which cancarry significant risk. Equity theory is also very relevant because it can be applied to just aboutany pay decision, because fairness is always a key concern.Moving away from individual-level theories, life cycle and diversification-based

contingency theories suggest that pay strategies should fit with corporate strategies. Theevolving empirical literature provides tentative support for many of the specific propositions.PAY PROGRAMSTable 2 summarizes the key features of some of the most widely used pay programs.Key dimensions include the payment method (whether increases roll into base salary or arepaid as bonuses or equity), the frequency of payouts, the nature of the performance measure,and who is typically covered under the different plans.Table 2. Comparison of Different Pay ProgramsIndividualIncentivesMerit Pay Merit Bonus Gainshairing Profit Sharing Ownership Skill BasedPayPayment

MethodBonus Changes inbase payBonus Bonus Bonus EquityChangesChanges inbase payPayoutFrequencyWeekly Annually Annually Monthly or



quarterlySemi-annuallyor annuallyWhen stocksoldWhen skillrequiredPerformanceMeasurementOutput,productivity,salesPerformanceratingPerformanceratingProduction or controllablecostsProfit Stock value Skill

acquisitionCoverage Direct labor AllemployeesAllemployeesProduction or service unitTotalorganizationTotalorganizationAllemployees

Source: Adapted and extended from Lawler, E.E. III. (1989). Pay for performance: A strategic analysis. In L.R. Gomez-Mejia(Ed.), Compensation and benefits. Washington, D.C.: Bureau of National Affairs.Employee Compensation WP 95-04Page 12

In compensating employees, an organization does not have to choose one program over another. Instead, a combination of programs is often the best solution. For example, oneprogram may foster teamwork and cooperation but not enough individual initiative. Another maydo the opposite. Used in conjunction, a balance may be attained. We now turn to a discussionof some recent trends in pay and an evaluation of .where such trends are likely to lead us.Recent DevelopmentsThe Shift to Variable Pay

According to a survey of over 2,000 U.S. companies by Hewitt Associates (Tully, 1993),the percentage of companies having a variable pay policy covering all salaried employeesincreased from 47% in 1988 to 68 in 1993. Moreover, whereas the standard merit increase(which rolls into base salary) was larger (5% versus 3.9%) in 1988 than the merit bonus (a lumpsum payment that does become part of base salary), by 1993 the situation was reversed withthe merit bonus being larger on average than the standard merit increase (5.9% versus 4.3%).Those in the human resource management field expect the movement toward variablepay to continue. In the Workplace 2000 study conducted by Dyer and Blancero (1993), 57human resource executives, consultants, academics, and others were asked to describe how



the workplace was likely to change by the year 2000. Dyer and Blancero provided studyparticipants the characteristics of a hypothetical service organization in 1991 and asked how itwould look in the year 2000. One expectation of participants was that pay would become morevariable. As Table 3 indicates, variable pay as a percentage of total direct compensation wasexpected to increase significantly for each of the four occupational groups studied.Table 4 provides some examples of how variable pay programs operate.

Table 3. Variable Pay as a Percentage of Total Direct Compensation1991 Scenario and Year 2000 ProjectionOccupational Group 1991 2000 Percentage ChangeExecutives 20 % 33 % 65 %Managers 10 % 23 % 130 %Professional/Technical 10 % 18 % 80 %Support 10 % 14 % 40 %Source: Dyer, L. & Blancero, D. (1993). Workplace 2000: Adelphi study. Center for Advanced HumanResource Studies, Cornell University.Employee Compensation WP 95-04Page 13

Table 4. Examples of Variable Pay Programs for Managers

Company Plan Participants Base Pay Policy Bonus PolicyNucor Steel 14 plant managers $80,000 to $150,000 (25 %below market)5 % of every dollar earnedbeyond 10 % return on equitygoes into bonus pool. Lastyear, average plant manager bonus equaled base salaryGeneral Mills Managers (Marketing manager in this example)$75,000 (versus $90,000market midpoint)

$10,000 if profits growth andreturn on capital are at marketaverage; up to $40,500 if profits growth and return oncapital are in top 10 % of market

AT&T 80,000 middle managers&30,000 scientists, researchers,technical employeesBetween 1986 and 1989, payraises less than one-half of competitors--move from payleader to below midpoint1. Individual or team bonus--pool depends in part oncorporate net profitability (5to 15 % of base dependingon individual/teamperformance)2. Business unit net profitability(about 2 % of base or less



this year)3. Corporate net profitability (7to 11 % of base)Source: Tully, S. (1993, November 1). Your paycheck gets exciting. Fortune, p. 83+.Employee Compensation WP 95-04Page 14

Group and Organization-Based Variable PayDyer and Blancero (1993) also found a strong belief that, in the future, variable paywould be based to a lesser degree on individual performance and to a much greater degree onfirth, business unit, and work group performance (see Table 3). The examples in Table 4 areconsistent with this expectation. It should be noted, however, that despite these significantchanges, Dyer and Blancero found that individual performance is expected to remain as thesingle most important determinant of variable pay for all occupational groups.Why are organizations making greater use of variable pay, and why are they movingaway from an individual focus to more of a group and organization focus? Variable pay is seenas a way of both controlling costs (especially in the case of organization-wide plans) andre-directing employee behavior.Better cost control is expected to be gained by replacing standard merit increases with

merit bonuses that are linked to firm or business unit performance. Thus, when profits or stockreturns are good, they can be shared with employees. However, when profits or stock returnsare poor or nonexistent, the organization is not saddled to the same degree with high fixed labor costs.In theory, the use of variable pay plans to control labor costs is fine and it even works inpractice under the right conditions; namely, if employees see a compelling business need tostay competitive in this manner. However, as in the widely discussed case of the DuPont Fibersdivision variable pay plan (Santora, 1991),3 employee opposition to downside variability in paywhen profit targets are not met can lead to such plans being discontinued as soon as the labor cost control aspect is supposed to kick in, and employees forego bonuses and receive only their (below market) base salary. This result is consistent with agency theory's prediction thatoutcome-oriented contracts are less successful when there is high outcome uncertainty.

Some organizations seek to avoid this "problem" by setting base pay at a higher level,and then sharing profits or stock with employees on top of their base salary during good years.These "gravy" plans do not control labor costs and, in fact, raise them. Yet, unless there is acompelling reason to believe that such pay plans significantly raise employee or for organization3 Under the DuPont plan, base salary was about 4 percent lower than for similar employees in other divisions, unless100 percent of the profit goal (a 4 percent increase over the previous year's profits) was reached. However, if the profitgoal wasexceeded, employees would earn more than similar employees in other divisions. For example, if the division reached150percent of the profit goal (i.e., 6 percent growth in profits), employees would receive 12 percent more than comparableemployees in other divisions. In 1989, when the profit goal was exceeded, the plan seemed to work fine. However, in1990,profits were down 26 percent from 1989, the profit goal was not met, and employees received no profit-sharing bonus.

Instead,they earned 4 percent less than comparable employees in other divisions. Employees were not happy and DuPonteliminatedthe plan and returned to a system of fixed base salaries with no variable component.Employee Compensation WP 95-04Page 15

productivity, organizations following this approach run the risk of investing extra money in theform of labor costs without realizing any return on the investment. Therefore, consistent withagency theory, employees may demand a compensating pay premium to assume risk. So, pay



risk costs the organization more money, but gains in effectiveness are not certain.Organizations that use variable pay, with or without downside risk, often believe thatsuch plans do generate significant returns. In agency theory terms, profit-sharing, stock plans,and gainsharing are examples of outcome-based contracts that seek to align the interests of employees and management with those of owners. As such, they are expected to re-directbehavior away from parochial individual goals, and more toward what it takes in terms of

cooperation, commitment, and innovation to make the group, business unit, or organization asuccess.

A change to variable pay may be a way to send a message to employees that things aregoing to change in important ways and therefore, may be helpful in supporting other major human resource changes. For example, variable pay may support a move to a team-basedorganization. As another example, variable pay may help eliminate the "entitlement" mentality or culture that can result from so-called merit increase plans that (in fact) fail to differentiatebetween employees with different performance levels, roll the increase into base salary so thecost remains in future years, and ignore the performance of the business. With a merit bonus,the pay has to be re-earned each year. Past individual performance does not matter, and is notreflected in base salary. Therefore, employees cannot rest on their past laurels. Moreover, thebonus pool may be linked to organization or business unit profitability. Again, the idea is to align

employee interests with those of the organization. In this case, the goal is to encouragecontinuous improvement and a forward-looking perspective.

Agency theory suggests that group and organization incentives can also contribute togreater overall levels of performance monitoring by, in effect, making each employee a principalwho monitors other employees (Levine & Tyson, 1990). So, if your pay and my pay depend onwhat we do as a team, we will be more likely to monitor each other's performance and givefeedback to one another when performance needs improvement: Similarly, according to equitytheory, if a person feels that his or her inputs (e.g., work effort) are greater than another member of the work group, but they receive the same reward, one way to restore equity wouldbe to encourage (or pressure) the other person to put forth more effort.Group size, however, is a key contingency variable in discussions of the behavioralimpact of group and organization variable pay plans. According to expectancy theory, the larger

the number of employees covered by a pay plan, the weaker the link they see between their Employee Compensation WP 95-04Page 16

own performance and pay (Schwab, 1973), and thus the weaker is their motivation. Similarly, atheme from the shirking, social loafing, and free rider literatures is that individual effortdecreases as the size of the group increases (Kidwell & Bennett, 1993).The implication, therefore, is that the ability of group and organization plans to changeemployee behavior may be very limited in cases where large numbers of employees arecovered. On this dimension, gainsharing plans, which typically cover smaller groups of employees, probably have an advantage over organization-wide plans like profit sharing andstock-based plans. Another advantage is that the performance measures in gainsharing plans(e.g., labor costs, quality) are often more controllable, again fostering greater employee

motivation to change behavior.The trade-off, however, is that gainsharing plans can pay off big even when thecompany is losing money. Another difficult situation arises when management would like tobring more work into the plant, but cannot afford to because the plan payouts would become toocostly. In these cases, one might say that gainsharing plans (consistent with the general historyof incentive plans) sometimes "fail" because they are too "successful." The payouts of anyincentive plan must walk the fine line between being too low to motivate employees and beingtoo high for management to afford. Even when standards work well initially, changes inproduction level and technology often result in the plan being unacceptable to one party or the



other. In some cases, management may choose to "buy out" employees by paying a lump sumsettlement in exchange for being able to redesign the plan with different standards, especially inunionized settings. An implication is that any sort of variable pay program should have a"sunset" provision that requires evaluation of the plan after a specific number of years, to avoidhaving the pay program becoming irrelevant because the organization changed, but it did not.

A final reason we discuss for the growth in variable pay plans is that the increased use

of total quality management (TQM) often entails a movement toward a team-based organizationand empowerment of employees to go beyond their traditional roles to make decisions in abroader range of areas that are likely to have an impact on organization performance.Individual-oriented systems may not be adequate for encouraging employees to pursue broadorganization goals, and to engage in the cooperative team and group-based decision-makingnecessary.

A survey conducted for the American Compensation Association (ACA) askedorganizations that implemented TQM programs how their pay practices changed (Davis, 1993).

As Table 5 indicates, major changes included less reliance on supervisors as the only source of Employee Compensation WP 95-04Page 17

performance appraisals, more reliance on team and organization results in setting pay, greater

use of variable pay, and fewer, broader pay grades.Table 5. Changes in Pay to Support Total QualityBefore After Performance Appraisal (n = 91)Only supervisors as source 59 % 12 %Peer/team appraisals 2 % 25 %

Add quality criteria/goals -- 68 %Have team goals -- 41 %Plan Increase Policies (n = 38)Increases tied to individual performance appraisals 88 % 60 %Increases tied to team/organization results 8 % 60 %Increases tied to quality results -- 49 %

Increases tied to skill/knowledge levels -- 33 %Incentive Program Policies (n = 56)Incentives based on individual results 26 % 31 %Incentives based on individual/ team results 23 % 37 %Incentives based on team/organization results 20 % 52 %Salary Structure PoliciesOther (e.g., more pay at risk) -- 52 %Fewer grades, broader range widths -- 38 %Note: n refers to the number of organizations (out of 196 total) that made changes in each pay area.Source: Davis, J.H. (1993, Autumn). ACA Journal. "Quality Management and Compensation."From an equity theory perspective, placing the entire employee population on such plansmay also create a greater sense of fairness among non-executive employees who typically have

not been covered by such plans in the past, but saw that executives were. Of course, this effectmay be limited to plans where variable pay is used to provide additional upside earningspotential, as opposed to cases where it replaces a portion of base salary.Banding, De-Layering, and Paying the Person Rather than the JobIn the traditional pay system, the worth of jobs is assessed on the basis of job evaluationdata in combination with market survey data. Job evaluation focuses on measuring and valuingEmployee Compensation WP 95-04Page 18

the specific characteristics and requirements of the job. Critics, however, suggest that job-based



systems tend to spawn too much bureaucracy, too much emphasis by employees on doing onlywhat is in their job description, and a lack of focus on market comparisons, which are critical for competitiveness. In addition, job levels become status indicators, which can get in the way. For example, an employee may be reluctant to accept a temporary assignment, that would be goodfrom a developmental point of view, unless it has at least as high of a job level.There have been at least two types of responses. First, organizations like General

Electric have cut levels of management and the corresponding pay grades. The goals are toimprove communication and speed decisions by reducing the levels of management, and toprovide wider pay grades (or bands) in order to allow more flexibility to recognize individualcontributions, and to make lateral movements simpler by reducing the likelihood of a job beingin a different (in this case, lower) grade (and looking like a demotion).The participants in the Dyer and Blancero (1993) study were also asked how the number of pay levels in the hypothetical service organization would change by the year 2000. Across thefour occupational groups, the 36 pay levels in 1991 were expected to decrease to 23 pay levelsby 2000, a decrease of about one-fourth. Whether the hoped for advantages of delayering andbanding will offset the potential drawbacks (e.g., less opportunity for promotion) remains to beseen.

Aside from allowing more flexibility in moving employees, banding, by virtue of a greater

spread between the minimum and maximum in each pay grade, is also intended to providemore opportunity to recognize individual differences in performance. So, within-level pay growthfor high performers will increase, while promotion opportunities and related pay growth willdecrease. It remains to be seen whether this will, on balance, be a good trade for motivationalpurposes. Further, banding carries the risk of becoming very expensive. Topping out of employees near the maximum would be very expensive under a banding system. Someorganizations have implemented sub-bands or zones within bands to avoid this problem.However, one might then reasonably ask what the difference is between an old system with 30grades and a new system with 10 bands, each with 3 sub-bands.

Another trend is for some organizations to move away from linking pay to job contentthrough job evaluation, and instead pay workers for the skills they possess. Skill-based pay linkspay to the breadth or depth of employee skill. The goal is to encourage learning, which in turn

facilitates flexibility in work assignments and encourages learning as a way of life to help withfuture organization change.Employee Compensation WP 95-04Page 19

Empirical EvidenceWhere are these recent developments in pay likely to lead us? We know that money canbe a powerful motivator. Indeed, a literature review of four motivational programs (individualmonetary incentives, goal-setting, job redesign, and participation in decision making) found thatmonetary incentives were associated with the largest average increase in physical productivity(Locke, Feren, McCaleb, Shaw, & Denny 1980). Therefore, changes in pay practices have thepotential to significantly change attitudes, behaviors, and organization functioning. Thechallenge, however, is to realize the potential of money as a motivator without running afoul of

the many roadblocks that arise in terms of measuring performance, setting standards that areperceived as fair, and choosing the right mix of individual, group, and organization objectives toreward.

As one recent example of a variable pay program gone wrong, consider the problemsSears encountered in some of its automotive repair shops in New Jersey and California. In aState of California undercover investigation, 38 visits to 27 Sears repair shops resulted in 34cases of unnecessary service or repair recommendations. Edward A. Brennan, the chairman of Sears, stated that "the incentive compensation program and sales goals created anenvironment where mistakes occurred" (Fisher, 1992). In essence, repair shop employees had



been rewarded for driving revenue (i.e., selling repairs to customers). Sears subsequentlychanged its pay system to one that focused on "quality."

Although specific examples are useful to demonstrate specific points, what does thebroader research literature tell us regarding the typical outcomes of variable pay and other payfor performance programs?

At the organization level, evidence suggests that greater emphasis on short-term

bonuses and long-term incentives (relative to base pay) is associated with higher subsequentprofitability, at least among top and middle level managers (Gerhart & Milkovich, 1990).Specifically, an organization with a bonus/base ratio of 10%, and 28% of its managers eligiblefor long term incentives had an average return on assets of 5.2%. In contrast, an organizationwith a 20% bonus to base ratio, and 48% of its managers eligible for long term incentives, hadan average return on assets of 7.1%.The fact that organization-based bonuses and incentives work for high-level managersdoes not necessarily mean they will work for other types of employees, most of whom have lessinfluence over organization performance and thus, weaker instrumentality perceptions. Still,even if the motivational impact (in terms of sheer effort) of organization-based incentives isEmployee Compensation WP 95-04Page 20

weaker for such groups, cost control and a re-focusing of behavior toward broader organizational goals may still be possible with such programs.The empirical evidence on profit sharing plans, in fact, generally paints a positive picture,with organizations using profit sharing having higher productivity (usually defined as valueadded per employee) on average than organizations that do not use profit sharing (e.g.,Weitzman & Kruse, 1990; Kruse, 1993a, 1993b). Still, there has yet to be a convincingdemonstration that profit sharing actually causes better organization performance (Gerhart &Milkovich, 1992). It may be that organizations with higher profit levels are more likely to adoptprofit sharing plans. In addition, if a profit sharing plan does not work out, it is likely to bediscontinued. So, the only profit sharing plans that are studied are those that have provensuccessful, and we do not hear about the plans that failed or needed to be replaced after theyserved their purpose. An organization that is deciding whether to adopt a profit sharing plan

must know how often such plans work and how often they fail or get discontinued, not just howwell the successful plans work.The evidence on stock plans is very limited, aside from the Gerhart and Milkovich (1990)study of top and middle level managers. The evidence that is available pertains mostly toemployee stock ownership plans (ESOPs). Like profit sharing, the evidence is generallyfavorable (Jones & Takao, 1993; Conte & Svejnar, 1990), but the same cautions regardingcausality apply. In any case, research suggests any beneficial effects of ESOPs may bestronger where employees have greater participation in making decisions, perhaps because itgives the employee a stronger feeling of ownership (Pierce, Rubinfeld, & Morgan, 1991). Thecosts of stock plans, especially options, may not always be obvious, but purchasing stock or issuing new stock (and the resulting dilution of the value of other shares) are costly moves.Indeed, U.S. Senator Carl Levin of Michigan introduced a bill in January 1993 that would require

companies to show the granting of stock options as an expense. The Financial AccountingStandards Board (FASB) also has proposed changes to its rules in this area.What is the evidence on gainsharing programs? Again, it is generally positive. Althoughthe types of cautions cited above regarding causality apply, the fact that from a theoreticalstandpoint, gainsharing programs offer employees a better line of sight (or instrumentality)between their performance and rewards (Lawler, 1989; Schwab, 1973) suggests that themotivational impact of such programs may be stronger than is the case with organization-wideprograms, like profit sharing and stock plans. Gainsharing payouts are typically based onmeasures like value added, sales value of production, or hours saved, which are more



controllable by employees than profits or stock performance.Employee Compensation WP 95-04Page 21

A time series study by Schuster (1983) of six gainsharing plans found substantial(around 30 %-) increases in productivity in four cases following the implementation of gainsharing. A fourth plant had an initial increase in productivity, but increases in the costs of

raw materials subsequently decreased the value added per worker, leading to no bonuses. Afifth plant, although not showing an increase in productivity, had gainsharing in place for twentyyears, suggesting that the plan was working, but the productivity increase had already occurredbefore the study. Other studies have also found significant productivity improvements fromgainsharing programs (e.g., Kaufman, 1992).In addition to having a payout measure that is controllable, gainsharing plans often havethe advantage of covering a smaller number of employees, which is also beneficial for motivation, because there is less likelihood of employees "free riding" (i.e., working less hardbecause others will work hard). Indeed, one study estimated that a doubling of employeescovered by a gainsharing plan from around 200 to 400 would reduce the expected productivitygain by almost one-half (Kaufman, 1992). The implication is that the number of employeescovered can have a substantial impact on the plan's success.

The fact that gainsharing (or any pay program) has a positive impact on productivity isno guarantee that it will continue to be used. A study of a gainsharing plan at an electrical utilityestimated a net savings of between $857,000 and $2 million, but the plan was discontinuedbecause employees in other divisions (all represented by the same union) felt unfairly treatedbecause they were not covered (Petty, Singleton, & Connell, 1992). The organization was thenfaced with two difficult options. First, it could include all employees under the same plan, butthat would likely increase the free rider problem and reduce the motivational impact. Second, itcould have a separate plan for each division, but this could easily result in unequal payoffs toemployees in different divisions, raising the same problems originally encountered withemployees and the union. There would also need to be a means of preventing between-divisioncompetition. A profit sharing or stock plan combined with gainsharing plans would be oneoption.

Other evidence also indicates that plans which appear to save money do not necessarilysurvive very long. Kaufman (1992) found that discontinued plans had improved productivitynearly as much as continuing plans. A study sponsored by the American Compensation

Association (McAdams & Hawk, 1992) may shed some light on this question. They found thatgainsharing plans, on average, were associated with net gains per employee of between $1,300and $3,700 per year. Nevertheless, when asked to rate the effectiveness of gainsharing plans inimproving effectiveness in areas like business performance, fostering teamwork, strengtheningEmployee Compensation WP 95-04Page 22

the pay-for-performance link, and so forth, the average effectiveness ratings all fell between2.63 and 3.25 on a 1 (no effectiveness) to 5 (high effectiveness) scale. In other words, mostrespondents were pretty lukewarm about gainsharing.Many organizations are moving to group and organization variable pay plans becausethey are frustrated with what they see as the failure of more traditional merit pay plans.Commonly cited problems include a lack of adequate differentiation between good and poor performers, employee and supervisor resistance, and the fact that merit increases sometimesseem to have become viewed as an entitlement by employees that is costly, and does not varywith business performance.

Although there is truth to many of these assertions, one sometimes wonders if perhapsmerit pay has been pronounced dead too soon. So-called studies of merit pay have often hadsignificant limitations (see Gerhart & Milkovich, 1992). In addition, the notion that there is no



merit pay is open to question.It is common to conclude that there is no individual merit pay because raises received bygood and poor performers differ by only a few percentage points. Two employees, each with abase salary of $40,000, one receiving a 5% increase, the other 6%, would receive raisesdiffering by $400 per year before taxes, or about $8 per week. Framed this way, the differencedoes indeed seem small and unlikely to motivate performance.

On the other hand, the example ignores the fact that high performers are more likely tobe promoted and thus, will have greater earnings growth. This is part of pay-for-performance,but it may not always be communicated as well as it could. Further, even limiting one's attentionto the annual increase process, it can be shown that small differences in pay raises accumulateinto significant differences over time. As Table 6 shows, the present value (or "real" payoff) toraises higher by 1 percentage point adds up to about $76,000 over 20 years. Factoring inpromotion based on performance and pay-linked benefits (e.g., retirement) would further increase the payoff to higher performers. Factoring in taxes would decrease the payoff. Wouldcommunicating the payoff to performance in this manner change the way employees react tomerit pay? Our conversations with managers yields a wide array of opinions on the matter,suggesting a good area for future research.Employee Compensation WP 95-04

Page 23Table 6. Pay for Performance: Accumulation over TimeEmployee 1 Employee 2Performance Average 1 point aboveRating above average

Annual Pay 5% 6%GrowthYear Nominal Real* Nominal Real'*1 $40,000 $40,0002 $42,000 $40,000 $42,400 $40,3813 $44,100 $40,000 $44,944 $40,7664 $46,305 $40,000 $47,641 $41,154

5 $48,620 $40,000 $50,499 $41,5466 $51,051 $40,000 $53,529 $41,9417 $53,604 $40,000 $56,741 $42,3418 $56,284 $40,000 $60,145 $42,7449 $59,098 $40,000 $63,754 $43,15110 $62,053 $40,000 $67,579 $43,56211 $65,156 $40,000 $71,634 $43,97712 $68,414 $40,000 $75,932 $44,39613 $71,834 $40,000 $80,488 $44,81914 $75,426 $40,000 $85,317 $45,24515 $79,197 $40,000 $90,436 $45,67616 $83,157 $40,000 $95,862 $46,111

17 $87,315 $40,000 $101,614 $46,55118 $91,681 $40,000 $107,711 $46,99419 $96,265 $40,000 $114,174 $47,44120 $101,078 $40,000 $121,024 $47,893Total $1,322,638 $760,000 $1,471,424 $ 836,690Difference $ 148,785 $ 76,690*Using 5 % discount rate.Employee Compensation WP 95-04Page 24



The empirical evidence on the effects of banding is basically non-existent. Research onskill-based pay is just beginning to emerge. An ACA survey of organizations using skill-basedpay illustrates some of the potential advantages and disadvantages. Most survey respondentsfelt that skill-based pay was successful in contributing to greater workforce flexibility andadaptability and in supporting work teams. However, relatively few saw any reduction in labor costs or layoffs. (Indeed, skill based pay is thought to permit a leaner headcount because of

cross-training.) So, one must consider whether possible higher labor costs are justified by theadvantages having to do with flexibility, adaptability, and the use of teams. Further, it must berecognized that if a plan is implemented, there are several factors that can contribute to itstermination. The ACA survey found the following to be most important in terminating skill basedpay plans: inadequate management commitment, unwillingness to endure short-termimplementation problems, poor plan designs that increase labor costs without providingoffsetting organizational benefits, conflicts between employees included and those excludedfrom the plan, inadequate training opportunities, and the failure of management to requiremeaningful skill certifications prior to pay increases.Only one skill based pay study to date (Murray & Gerhart, 1994) has used objectivemeasures of productivity and quality, a control group, and a time series before and after implementation of the plan. In a comparison of two automobile parts plants, Murray and Gerhart

found that a significant increase in productivity and product quality took place in the plant thatimplemented skill based pay.Globalization and CompensationThe continued globalization of markets means that we will have to increasingly consider whether the effect of different pay strategies is likely to differ from country to country, or between cultures within a country. Hofstede's (1993) work on identifying culture differences ondimensions such as power distance (i.e., the degree of inequality considered normal),individualism, masculinity, uncertainty avoidance; and short- versus long-term time orientationhas been used by Hodgetts and Luthans (1993) to begin studying this question. Certainhypotheses flow readily from the national differences depicted in Table 7.4 Variable pay (pay atrisk) may face difficulties in countries that have a high need for uncertainty avoidance such asJapan, South Korea, and Taiwan. Individualistic programs such as merit pay could be a problem

in cultures where collectivism is a stronger norm than individualism (e.g., the Pacific Rimcountries). Still, average differences in culture are just that, averages, and should not4 Long- versus short-term orientation is not shown. Japan, Hong Kong, and China have a long-term orientation,whereas the United States has a more short-term oriented culture.Employee Compensation WP 95-04Page 25

necessarily be viewed as factors that must be taken as a given. Honda, in Japan, for example, just recently announced that it would be changing many of its managers over to a merit paysystem.On the other hand, U.S. companies that have attempted to export pay practicesoverseas have often encountered difficulties. Lincoln Electric, famous for its history of successusing variable pay, has thus far not been successful in implementing variable pay in itsoverseas acquisitions. Our own experience with gainsharing plans in Western Europe has notbeen successful (Chilton, 1993). The cultural differences described by Hofstede and relatedcustoms are often difficult to overcome. It is probably significantly easier to implement paypractices that are not typical of a country in a greenfield setting as opposed to an acquisition.Indeed, Japanese (e.g., Honda, Nissan) and German (e.g., BMW,Mercedes-Benz) automobileplants opened in the United States have often been in greenfield sites, where the company hasmaximum flexibility in screening and choosing employees who will fit well with their corporateculture, human resource management, and pay philosophies.Table 7. National Culture Clusters



Region or Country Power DistanceIndividualism Masculinity Uncertainty

AvoidancePacific RimHong Kong, Malaysia,

Philippines, SingaporeHigh Low High LowJapan High Low High HighSouth Korea, Taiwan High Low Low HighUnited States, Great Britain Low High High LowSources:Hofstede, G. (1993, February). Cultural constraints in management theories. Academy of Management Executive, 7(1), 81-94; Hodgetts, R.M. & Luthans, F. (1993, March-April). U.S.multinationals' compensation strategies for local management: Cross-cultural implications.Compensation & Benefits Review, 25, 42-48. Review, 25, 42-48.Employee Compensation WP 95-04Page 26

CONCLUSION

Our goal in this chapter has been to describe the theory and practice of compensation,as well as provide an overview of recent empirical evidence on the consequences of differentcompensation practices. The theory section points to the many trade-offs in designing employeecompensation policies. Examples of trade-offs include maximizing high individual effort versusteamwork and cooperation, controlling costs versus maximizing employee effort, and providingincentives for promotion versus producing feelings of inequity due to large pay differentials. Our message has been that the nature of such trade-offs should depend on the corporate andbusiness strategies and that the trade-offs can be made less of a problem by combining payprograms in a way that helps balance competing objectives.

As a final comment, we would like to emphasize that, although it is important to keepabreast of what other organizations are doing (benchmarking) in the area of employeecompensation, it is crucial to remember that what works for one organization may not work at all

for another. Therefore, surveys of "best practices" are useful to the extent that the surveysreport a diversity of best practices and the reasons why different practices are best for differentorganizations. The ultimate choice of a best compensation strategy rests, of course, on its fitwith other human resource activities and its fit with the business strategy.Employee Compensation WP 95-04Page 27

REFERENCESAdams, J.S. (1963). Toward an understanding of inequity. Journal of Abnormal Psychology, 67, 422-436.Beam, B.T. Jr., & McFadden, J.J. (1992). Employee benefits. Chicago: Dearborn Financial Publishing.Bretz, R.D., Ash, R.A., & Dreher, G.F. (1989). Do people make the place? An examination of theattraction-selection-attrition hypothesis. Personnel Psychology, 42, 561-581.Chilton, K. (1993, November-December). Lincoln Electric's incentive system: Can it be transferred overseas?Compensation and Benefits Review, 21-30.Conte, M.A., & Svejnar, J. (1990). The performance effects of employee ownership plans. In A.S. Blinder (Ed.),Paying for productivity, pp. 245294. Washington, D.C.: The Brookings Institution.Cowherd, D.M., & Levine, D.I. (1992). Product quality and pay equity between lower-level employees and topmanagement: An investigation of distributive justice theory. Administrative Science Quarterly, 37, 302-320.Davis, J.H. (1993, Autumn). ACA Journal. Quality management and compensation.Dyer, L., & Blancero, D. (1993). Workplace 2000: A delphi study. Center for Advanced Human ResourceStudies,Cornell University.



Eisenhardt, K.M. (1989). Agency theory: An assessment and review. Academy of Management Review, 14,57-74.Ellig, B.R. (1981). Compensation elements: Market phase determines the mix. Compensation Review, (ThirdQuarter), 30-38.Fama, E.F., & Jensen, M.C. (1983). Separation of ownership and control. Journal of Law and Economics, 26,301-325.

Fisher, L.M. (1992, June 23). Sears auto centers to halt commissions. New York Times, D1.Gerhart, B., & Milkovich, G.T. (1990). Organizational differences in managerial compensation and financialperformance. Academy of-Management Journal, 33, 663-691.Gerhart, B., & Milkovich, G.T. (1992). Employee compensation: Research and practice. In M.D. Dunnette , &L.M.Hough (Eds.), Handbook of industrial and organizational Psychology, 2nd Edition, pp. 481-569. Palo Alto,CA: Consulting Psychologists Press, Inc.Gerhart, B., Milkovich, G.T., & Murray, B. (1992). Pay, performance, and participation. In D. Lewin, O. Mitchell,& P.Sherer (Eds.), Research Frontiers in Industrial Relations, pp. 193-238. Madison, WI: Industrial RelationsResearch Association.Gomez-Mejia, L.R. (1992). Structure and process of diversification, compensation strategy, and firm

performance.Strategic Management Journal, 13, 381-397.Gomez-Mejia, L.R., & Balkin, D.B. (1992). Compensation organizational strategy and firm performance.Cincinnati:South-Western.Employee Compensation WP 95-04Page 28

Greenberg, J. (1990). Employee theft as a reaction to underpayment of inequity: The hidden cost of pay cuts.Journalof Applied Psychology, 75, 561-568.Heneman, H.G., III , & Schwab, D.P. (1979). Work and rewards theory. In D. Yoder , & H.G. Heneman, Jr.(Eds.),A

SPA

handbook of personnel and industrial relations. Washington, D.C.: Bureau of NationalA

ffairs.Hodgetts, R.M., & Luthans, F. (1993, March-April). U.S. multinationals' compensation strategies for localmanagement: Cross-cultural implications. Compensation & Benefits Review, 25, 42-48.Hofstede, G. (1993, February). Cultural constraints in management theories. Academy of ManagementExecutive,7(1), 81-94.Jones, D.C., & Takao, K. (1993). The scope, nature, and effects of employee stock ownership plans in Japan.Industrial and Labor Relations Review, 46, 352-367.Judge, T.A., & Bretz, R.D. Jr. (1992). Effect of values on job choice decisions. Journal of Applied Psychology,77,261-271.Kaufman, R.T. (1992). The effects of Improshare on productivity.- Industrial and Labor Relations Review, 45,

311-322.Kerr, J.L. (1985). Diversification strategies and managerial rewards. Academy of Management Journal, 28,155-179.Kidwell, R.E. Jr., & Bennett, N. (1993). Employee propensity to withhold effort: A conceptual model tointersect threeavenues of research. Academy of Management Review, 18, 429-456.Kruse, D.L. (1993a). Profit sharing: Does it make a difference? Kalamazoo, MI: Upjohn Institute.Kruse, D.L. (1993b). Does profit sharing affect productivity? Tests using panel data on profit sharing featuresand



personnel policies. Working Paper, Rutgers University.Lambert, R.A., & Larcker, D.F. (1989). Executive compensation, corporate decision-making, and shareholder wealth.In F. Foulkes (Ed.), Executive compensation (pp. 287-309). Boston: Harvard Business School Press.Lawler, E.E. III. (1989). Pay for performance: A strategic analysis. In L.R. Gomez-Mejia (Ed.), Compensationand

benefits. Washington, D.C.: Bureau of National Affairs.Levine, David I., & Tyson, Laura D. (1990). Participation, productivity, and the firm's environment. In Alan S.Blinder (Ed.), Paving for productivity. Washington, D.C.: Brookings Institution.Locke, E.A., Feren, D.B., McCaleb, V.M., Shaw, K.N., & Denny, A.T. (1980). The relative effectiveness of four methods of motivating employee performance. In K.D. Duncan, M.M. Gruenberg, & D. Wallis (Eds.),Changes in working life (pp. 363-388). New York: Wiley.McAdams, J.L., & Hawk, E.J. (1992). Capitalizing on human assets through performance-based rewards. ACA Journal, 1(1), 60-73.Miles, R.E., & Snow, C.C. (1978). Organizational strategy, structure, and process. New York: McGraw-Hill.Milkovich, G. T. (1988). A strategic perspective on compensation management. Research in Personnel andHuman

Resources Management, 6, 263-288.Employee Compensation WP 95-04Page 29

Milkovich, G.T., Gerhart, B., & Hannon, J. (1991). The effects of research and development intensity onmanagerialcompensation in large organizations. Journal of High Technology Management Research, 2, 133-150.Milkovich, G.T., & Newman, J. (1993). Compensation (4th Ed.). BPI/Irwin: Homewood, IL.Milkovich, G.T., & Wigdor, A.K. (1991). Pay for performance. Washington, D.C.: National Academy Press.Murray, B., & Gerhart, B. (1993). Organizational outcomes from the introduction of a skill-based pay program.Working Paper. Center for Advanced Human Resource Studies, Cornell University.Noe, R.A., Hollenbeck, J.R., Gerhart, B., & Wright, P.M. (1994). Human resource management: Gaining acompetitive advantage. Burr Ridge, IL: Austen Press /Irwin.

Petty, M.M., Singleton, B., & Connell, D.W. (1992).A

n experimental evaluation of an organizational incentiveplan inthe electric utility industry. Journal of Applied Psychology, 77, 427-436.Pierce, J.L., Rubenfeld, S., & Morgan, S. (1991). Employee ownership: A conceptual model of process andeffects.Academy of Management Review, 16, 121-144.Santora, J.E. (1991, February). DuPont returns to the drawing board. Personnel Journal, 34-36.Schuster, M. (1983). The impact of union-management cooperation on productivity and employment.Industrial andLabor Relations Review, 36, 415-430.Schwab, D.P. (1973). Impact of alternative compensation systems on pay valence and instrumentalityperceptions.

Journal of A

pplied Psychology, 58, 308-312.Tosi, H.L. Jr., & Gomez-Mejia, L.R. (1989). The decoupling of CEO pay and performance: An agency theoryperspective. Administrative Science Quarterly, 34, 169-189.Tully, S. (1993, November 1). "Your Paycheck Gets Exciting." Fortune, p. 83+.Vroom, V.H. (1964). Work and motivation. New York: Wiley.Weitzman, M.L., & Kruse, D.L. (1990). Profit sharing and productivity. In A.S. Blinder (Ed.), Paying for productivity.Washington, D.C.: Brookings Institution.

Documents

Employee Compensation - Theory, Practice & Evidence